Tennis coaching system, apparatus, and operating method using artificial intelligence

Abstract

Embodiments of the present invention include a tennis coaching system using artificial intelligence, the tennis coaching system comprising: a terminal that obtains user posture measurement information and transmits the posture measurement information to a server; and a server that obtains a user posture video, calculates a similarity between the user posture video and a comparison video, and provides coaching information on the basis of the similarity.

Description

Tennis coaching system, device, and operating method using artificial intelligence

[0001] The present invention relates to a tennis coaching system, device and method using artificial intelligence.

[0002] Traditional tennis education and training methods have relied primarily on offline lessons. These are conducted individually or in small groups, with the availability of professional coaches and adequate training space being key factors.

[0003] However, the lack of professional coaches and space constraints in urban areas make it difficult for all tennis learners to receive timely coaching.

[0004] This poses a significant obstacle for individuals seeking to improve their tennis skills, leading to problems such as longer wait times and increased costs for learners to receive coaching.

[0005] Moreover, digital tennis games like screen tennis have recently gained popularity. While these systems allow users to enjoy tennis in a virtual environment, they are largely focused on gaming and entertainment, lacking specific services for improving tennis skills or coaching. This reflects the reality that, despite technological advancements, various digital tennis platforms, including screen tennis, fail to meet the needs for practical skill improvement and personalized coaching.

[0006] Against this backdrop, the need for a new method for tennis learners to receive professional coaching without the constraints of time and place is emerging.

[0007] The purpose of this disclosure is to address the shortage of professional coaches and the difficulty in securing adequate training space by using an artificial intelligence-based tennis coaching system.

[0008] Furthermore, the purpose of this disclosure is to address the lack of practical coaching services not provided by existing play-oriented platforms such as screen tennis by using a tennis coaching system utilizing artificial intelligence.

[0009] In addition, the purpose of the present disclosure is to provide a tennis application system that enables a user to improve his / her skills without the guidance of a professional coach by providing an AI-based coaching function that can systematically analyze and improve the user's tennis skills.

[0010] According to an embodiment of the present invention, a tennis coaching system using artificial intelligence may include a terminal that obtains user posture measurement information and transmits the posture measurement information to a server, and a server that obtains a user posture image, calculates a similarity between the user posture image and a comparison target image, and provides coaching information based on the similarity.

[0011] Additionally, the comparison target image may include at least one of a coach image and a previous user posture image of the user.

[0012] In addition, the comparison target image may include one of the first user posture image having the highest similarity to the comparison target image among the user's previous user posture images, the first image among the user's posture images, and the previously used comparison target image.

[0013] In addition, the server can obtain a plurality of body joint points from the user posture image, determine a reference coordinate system for determining a reference of the number of body joint points, obtain coordinate values ​​of each of the plurality of body joint points based on the reference coordinate system, and obtain a similarity based on the coordinate values.

[0014] In addition, the server can obtain a first angle calculated as an angle between adjacent joints based on the coordinate values ​​of each joint point of the comparison target image, and a second angle calculated as an angle between adjacent joints based on the coordinate values ​​of each joint point of the user posture image, and can calculate the user's motion similarity based on the difference between the first angle and the second angle.

[0015] In addition, the terminal may output at least one screen among a posture measurement information input screen for obtaining the posture measurement information, a comparison target image selection screen for determining a comparison target image, and a guidance screen for providing a guidance guide while obtaining the posture measurement information, and transmit the posture measurement information and information about the comparison target image obtained based on user input to the server.

[0016] Additionally, the above-described posture measurement information may include at least one of posture information to be practiced, handle information, ball launch interval information, ball quality information, number of swings information, ball drop location information, and swing guide information.

[0017] In addition, the coaching information may include at least one of a similarity calculation result between the user posture image and the comparison target image, posture measurement information, the user posture image record, options of the comparison screen, and feedback information for the user posture image.

[0018] In addition, the method further includes a user terminal that receives the coaching information from the server and outputs the received coaching information, and the user terminal can output at least one of a comparison area that shows a similarity comparison result between the user posture image and the comparison target image, an information display area that includes posture measurement information, a recording area that records the user posture image, an option area that sets options for a comparison screen, and a feedback area that provides feedback on the user posture image.

[0019] In addition, the coaching information includes at least one of comprehensive information, detailed coaching information, posture accuracy information, posture consistency information, and consistency graph information, and the detailed coaching information can be determined based on the similarity of each of a plurality of detailed movements included in each of the user posture image and the comparison target image.

[0020] Additionally, the plurality of detailed movements can be determined based on the coordinate values ​​of wrist joint points obtained from the user posture image.

[0021] Additionally, the consistency graph can represent time information corresponding to the coordinate values ​​of the wrist joint points obtained from the user posture image.

[0022] In addition, the server according to an embodiment of the present invention may divide the user posture image and the comparison target image into a plurality of images based on posture motion, divide each of the plurality of images divided based on the posture motion into detailed motions constituting the posture motion, and calculate the similarity based on the difference between the first angle acquired at each joint point of the detailed motion of the comparison target image and the second angle acquired at each joint point of the detailed motion of the user posture image.

[0023] In addition, the server can obtain a plurality of user posture motion images segmented based on posture motion from the user posture image, identify a first user posture motion image as a reference among the plurality of user posture motion images, calculate the similarity between the first user posture motion image and other user posture motion images, and calculate consistency information based on the respective similarities.

[0024] In addition, the processor can determine the similarity for each detailed action, and if the similarity is determined to be below a set value, identify an image of the detailed action, and generate coaching information based on the identified image.

[0025] According to an embodiment of the present invention, by providing an artificial intelligence-based coaching system, a player can more effectively improve his or her tennis skills.

[0026] Additionally, according to embodiments of the present disclosure, users can continuously monitor their progress through an AI-based coaching system and receive systematic feedback tailored to their individual growth goals, thereby facilitating self-directed learning and practice.

[0027] Figure 1 illustrates a tennis coaching system (1) according to an embodiment of the present invention.

[0028] Figure 2 is a block diagram showing the configuration of a tennis coaching system (1) according to an embodiment of the present invention.

[0029] Figure 3 is a flowchart showing the operation of a terminal (200) according to an embodiment of the present invention.

[0030] Figure 4 illustrates a detail measurement information input screen according to an embodiment of the present invention.

[0031] Figure 5 illustrates a comparison target image selection screen according to an embodiment of the present invention.

[0032] Figure 6 illustrates a progress guidance screen according to an embodiment of the present invention.

[0033] Figure 7 illustrates an example of a user posture image according to an embodiment of the present invention.

[0034] Figure 8 illustrates a waiting guidance screen output after acquiring a user posture image according to an embodiment of the present invention.

[0035] Figure 9 is a flowchart showing the operation method of a server for providing an AI coaching service according to an embodiment of the present invention.

[0036] FIG. 10 is a conceptual diagram illustrating a method for a server according to an embodiment of the present invention to obtain coordinates of joint points of a human body.

[0037] FIG. 11 is a flowchart illustrating a method for calculating the similarity of user posture images according to an embodiment of the present invention.

[0038] FIG. 12 is a diagram illustrating an example of posture similarity provided by a server to a user according to an embodiment of the present invention.

[0039] FIG. 13 is a flowchart illustrating a method for determining consistency information from a user posture image according to an embodiment of the present invention.

[0040] FIG. 14 is a flowchart illustrating a method by which a server determines a user posture image for determining similarity according to an embodiment of the present invention.

[0041] Figure 15 illustrates a coaching screen for outputting coaching information according to an embodiment of the present invention.

[0042] FIG. 16 illustrates an embodiment of a feedback area (1550) for outputting coaching information according to an embodiment of the present invention.

[0043] FIG. 17 illustrates an embodiment of a feedback area (1550) for outputting coaching information according to an embodiment of the present invention.

[0044] Figures 18 and 19 illustrate examples of consistency graphs based on a forehand motion according to an embodiment of the present invention.

[0045] Figure 20 illustrates the configuration of a terminal, server, and user terminal of a tennis coaching system (1) according to an embodiment of the present invention.

[0046] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed in this specification are merely illustrative for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described in this specification.

[0047] Embodiments according to the concept of the present invention may have various modifications and take various forms, and thus, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit embodiments according to the concept of the present invention to specific disclosed forms, but rather includes all modifications, equivalents, or alternatives included within the spirit and technical scope of the present invention.

[0048] While terms such as "first" or "second" may be used to describe various components, these components should not be limited by these terms. These terms are intended solely to distinguish one component from another. For example, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component, without departing from the scope of the invention.

[0049] The terminology used herein is for the purpose of describing specific embodiments only and is not intended to limit the present invention. The singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, it should be understood that the terms "comprises" or "has" indicate the presence of a described feature, number, step, operation, component, part, or combination thereof, but do not preclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0050] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0051] Also, the term "and / or" as used herein should be understood to refer to and include all possible combinations of one or more of the associated items listed. Furthermore, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified or clear from the context, "X utilizes A or B" means that X utilizes A; X utilizes B; or X utilizes both A and B, and "X utilizes A or B" can apply to any of these cases. Furthermore, "at least one selected from A and B" can refer to (1) A, (2) at least one of A, (3) B, (4) at least one of B, (5) at least one of A and at least one of B, (6) at least one of A and B, (7) at least one of B and A, or (8) both A and B.

[0052] As used herein, the singular expressions should be understood to include the plural expressions unless the context clearly dictates otherwise, and the term "comprises" and the like should be understood to mean the presence of a described feature, number, step, operation, component, part, or combination thereof, but not to exclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0053] Before going into a detailed description of the drawings, it should be made clear that the division of components in this specification is merely a division based on the main function of each component. In other words, two or more components described below may be combined into a single component, or a single component may be further subdivided into two or more components with more detailed functions. In addition to its own main function, each component described below may additionally perform some or all of the functions of other components, and of course, some of the main functions of each component may be exclusively performed by other components.

[0054] Additionally, in performing a method or method of operation, each process constituting the method may occur in a different order than the stated order, unless the context clearly indicates a specific order. That is, each process may occur in the same order as the stated order, may be performed substantially simultaneously, or may be performed in the opposite order.

[0055] Hereinafter, the present invention will be described in detail by describing a preferred embodiment of the present invention with reference to the attached drawings.

[0056] Tennis Coaching System

[0057] Figure 1 illustrates a tennis coaching system (1) according to an embodiment of the present invention.

[0058] A tennis coaching system (1) according to an embodiment of the present invention may include a server (100), a terminal (200), a ball machine device (300), a camera (400), and a user terminal (500) as main components for providing a tennis coaching service using artificial intelligence.

[0059] According to an embodiment of the present invention, the server (100) can analyze the user's posture measurement information input via the terminal (200) and the captured posture image of the user to provide coaching information. To this end, the server (100) may include an artificial intelligence model required for tennis coaching.

[0060] According to an embodiment of the present invention, a terminal (200) is an electronic device that interacts with a user who wishes to utilize a tennis coaching service. The terminal (200) can obtain posture measurement information, which is information about the posture for which the user wishes to receive coaching. The terminal (200) can transmit the posture measurement information to a server (100), a ball machine device (300), and a camera (400). The tennis coaching system (1) can obtain coaching information that the user wishes to receive through the coaching service based on the information obtained through the terminal (200).

[0061] According to an embodiment of the present invention, the ball machine device (300) is a ball supply device that supplies a ball based on the user's posture measurement information. The ball machine device can control the speed, angle, and timing of the ball to provide optimized training to the user based on information obtained from the terminal (200) or server (100).

[0062] According to an embodiment of the present invention, a camera (400) is provided to capture a user's posture and can capture the user's movements to obtain a user posture image. The user posture image can be used as image data for analyzing the user's tennis posture.

[0063] The above camera (400) may be a photographing device mounted on a specific facility or a camera attached to an electronic device such as a smartphone or tablet.

[0064] The above camera (400) can be positioned in an area of ​​a tennis court to capture the user's movements. Furthermore, the area where the filming device is positioned may be a predetermined area on the tennis court, and the predetermined area may vary based on the posture (movement) the user wishes to receive coaching for.

[0065] According to an embodiment of the present invention, a user terminal (500) receives analyzed tennis coaching information from a server. The user terminal (500) may be equipped with a display for displaying the coaching information. This allows the user to check and improve their performance.

[0066] According to an embodiment, the terminal (200) may perform the role of the user terminal (500), in which case the user terminal (500) may be omitted.

[0067] Meanwhile, according to one embodiment of the present disclosure, the server (100) and the terminal (200) can transmit and receive tennis coaching-related data through wireless and / or wired interconnection.

[0068] The above server and terminal can be implemented as a computing device, and the computing device is a device that processes input data on a regular basis and performs operations required for a tennis coaching system according to a specific model or algorithm.

[0069] For example, a computing device may be implemented in the form of a PC, an electronic device, a server on a network, a smart device, a chipset with an embedded design program, etc.

[0070] In this way, the tennis coaching system (1) according to an embodiment of the present invention provides tennis coaching information based on the similarity between the user posture image and the comparison target image, thereby enabling the user to improve his or her tennis skills.

[0071] Hereinafter, embodiments of the present invention will be specifically described based on the operation of each component.

[0072] Figure 2 is a block diagram showing the configuration of a tennis coaching system (1) according to an embodiment of the present invention.

[0073] Referring to FIG. 2, the terminal (200) can be connected to a camera (400), a user terminal (500), a server (100), and a ball machine device (300) to transmit and receive data.

[0074] According to an embodiment of the present invention, the terminal (200) can execute a tennis coaching service program utilizing artificial intelligence (S201). The tennis coaching service program can be implemented in the form of an application or webpage that users can easily access and use.

[0075] The above tennis coaching service program allows users to easily input posture measurement information for which they wish to receive coaching through a user interface and select options for the posture for which they wish to receive coaching.

[0076] First, the terminal (200) can obtain the user's posture measurement information (S202).

[0077] At this time, the posture measurement information may refer to the posture the user wishes to receive coaching for and options that must be set in advance to provide coaching services.

[0078] For example, the above-described posture measurement information may include at least one of posture information to be practiced (forehand, backhand, volley, serve, slice, etc.), grip information, ball launch interval information, ball quality information, number of swings information, ball drop location information, and swing guide information.

[0079] According to an embodiment, the above-described posture measurement information may be input by a user.

[0080] According to an embodiment of the present invention, the terminal (200) can obtain posture measurement information and determine a comparison target image (S203).

[0081] The above comparison target video may refer to an answer or guide video for reference when a user wants to receive coaching on a specific posture.

[0082] The above comparison target image may include at least one of a coach image and a previous user posture image of the user.

[0083] The above comparison target video can be selected from a demonstration video of a professional coach or a video of the user's previous posture.

[0084] Meanwhile, according to an embodiment of the present invention, the terminal (200) can output a usage guide screen for a usage guide of the tennis coaching service program.

[0085] The above user guide screen may be composed of at least one screen, and the screen may be designed so that the user can effectively use the tennis coaching service program.

[0086] Users will be able to determine the posture measurement information and comparison target images by interacting with the above user guide screen.

[0087] Specifically, the above-mentioned user guide screen may include at least one screen among a posture measurement information input screen for obtaining the posture measurement information, a comparison target image selection screen for determining a comparison target image, and a progress guidance screen for providing guidance while obtaining the posture measurement information.

[0088] As above, basic information for using a tennis coaching service using artificial intelligence was obtained using the terminal (200).

[0089] According to an embodiment of the present invention, the terminal (200) can transmit posture measurement information to at least one of a server (100), a ball machine device (300), and a camera (400).

[0090] In addition, the terminal (200) can transmit a control command for the ball machine device (300) to supply the ball. At this time, the control command may include a control signal for the user's desired posture, ball quality, speed, number of times, drop location, etc. based on the posture measurement information.

[0091] The above ball machine device (300) can supply a ball based on a control command received from the terminal (200) (S301).

[0092] Additionally, the camera (400) according to an embodiment of the present invention can receive a shooting control command from the terminal (200). The camera can capture the user's tennis posture in response to the received shooting control command (S401). The user's tennis posture thus acquired corresponds to a user posture image.

[0093] Meanwhile, according to an implementation example, the camera (400) may determine in real time a shooting position for obtaining a user posture image based on the received posture measurement information, and may also be able to shoot the user at the determined position.

[0094] Afterwards, the camera can transmit the user posture image to the server (100).

[0095] The server (100) can obtain a user posture image and analyze the user's posture using an artificial intelligence model (S101).

[0096] The server (100) according to an embodiment of the present invention can provide coaching information to a user based on the similarity of the posture measurement information and the comparison target image.

[0097] Specifically, user posture analysis can be determined based on the similarity between the comparison target image and the user posture image. For example, user posture analysis may include preprocessing the user posture image.

[0098] Additionally, user posture analysis can utilize a pose estimation model.

[0099] For this purpose, the above-mentioned posture estimation model may be a pre-learned artificial intelligence model.

[0100] Details related to this will be explained in detail later.

[0101] Thereafter, the server (100) may provide an artificial intelligence-based tennis coaching service by providing the above coaching information to a user terminal (500) or terminal (200) (S501).

[0102] Meanwhile, depending on the embodiment, a camera (400) may be provided in the terminal (200), and the user posture analysis may also be implemented as a program installed in the terminal (200).

[0103] Below, the process of determining posture measurement information and comparison target images using a terminal (200) and receiving coaching information thereafter is described.

[0104] Figure 3 is a flowchart showing the operation of a terminal (200) according to an embodiment of the present invention.

[0105] According to an embodiment of the present invention, the terminal (200) can obtain information necessary for a tennis coaching system by using the user guide screen.

[0106] Specifically, the terminal (200) can acquire user posture measurement information using the posture measurement information input screen (S310). As previously described, the posture measurement information may include the type of lesson movement to be practiced (e.g., forehand, backhand, volley, serve, etc.), grip method, ball launch interval, number of swings, ball landing location, etc.

[0107] The user will be able to input posture measurement information about his or her tennis posture through the terminal (200).

[0108] Meanwhile, according to various embodiments, the posture measurement information can also be set as follows based on the tennis coaching service usage history.

[0109] <If there is no usage history>

[0110] 1. Users directly enter posture measurement information through the terminal.

[0111] 2. Use of basic posture measurement information provided by the terminal (200)

[0112] 3. Use of posture measurement information received from the server or user terminal

[0113] <If there is a history of use>

[0114] 1. Set the same posture measurement information as the usage history

[0115] According to an embodiment of the present invention, the terminal (200) can output a comparison target image selection screen to provide a list of comparison target images to the user (S320).

[0116] The user can select an image to be compared with his / her own image from the list of images to be compared displayed on the image selection screen.

[0117] For example, the comparison target video may include either a demonstration video by a professional coach or a video of the user's previous posture. The comparison target video may be used as a reference video for comparison analysis with the user's current posture.

[0118] Below is a description of the types of images to be compared.

[0119] According to an embodiment of the present invention, the comparison target image may include a previous image captured by the tennis coaching system and stored in a user account.

[0120] This is a previous user posture video acquired by the user in a facility equipped with a tennis coaching system. The previous user posture video is a video previously stored in the tennis coaching system.

[0121] Users will be able to compare this video to their current skill level.

[0122] According to an embodiment of the present invention, the comparison target video may include a video stored in a user account after receiving coaching based on a tennis coaching system.

[0123] This is a previous user posture image acquired while the user was receiving coaching at a facility equipped with a tennis coaching system. For example, it may be the first user posture image with the highest similarity to the comparison target image.

[0124] The tennis coaching system can store and retain the user's previous posture videos in the user's account. Users can adjust their practice by accurately executing the coached techniques through the videos.

[0125] Alternatively, the comparison video may include a video taken directly by the user from the same angle as the tennis coaching system. The user may film their own play by imitating the angle of the tennis coaching system.

[0126] Users will be able to objectively analyze their skills even in environments where they do not use a tennis coaching system.

[0127] According to an embodiment of the present invention, the comparison target image may include a first posture image of the user acquired by the tennis coaching system.

[0128] In other words, users can set their own tennis play videos as comparison targets. These videos document the user's skill level and posture, and can serve as an important comparison point for evaluating the effectiveness of subsequent training and coaching.

[0129] According to an embodiment of the present invention, the comparison target image may include an image provided from a coach data list stored in a terminal or server.

[0130] The server or terminal (200) included in the tennis coaching system has a built-in coach data list for various tennis techniques and postures, and the user can select an appropriate coaching video from the coach data list.

[0131] The above videos can assist users in accurately understanding and practicing the skills they are targeting, including demonstrations by professional coaches.

[0132] Through these various types of comparative videos, users will be able to analyze their tennis play from various perspectives, and objectively compare and analyze their posture and techniques based on these videos.

[0133] According to an embodiment of the present invention, the terminal (200) can output a comparison target image selected by a user input and provide a guide for imitating the comparison target image.

[0134] According to an embodiment of the present invention, the terminal can transmit the posture measurement information to a ball machine device and control the camera to capture a user posture image (S330).

[0135] Specifically, according to an embodiment of the present invention, a tennis coaching system providing an AI coaching function provides posture measurement information to a ball machine device (300) and a camera (400) via a terminal (200), thereby enabling a user to obtain ball feeding and user posture images based on settings. The above process may be composed of the following detailed steps.

[0136] According to an embodiment of the present invention, the terminal (200) can generate control commands for the ball machine device (300) and the camera (400) based on posture measurement information input from the user. The control commands can be transmitted to the ball machine device and the camera via a wireless or wired network.

[0137] The ball machine device can obtain posture measurement information.

[0138] The posture measurement information transmitted from the above terminal will include detailed settings such as the posture to be practiced, grip, ball launch interval, pitch quality, number of times, ball drop location, and spin amount according to the user's practice goal.

[0139] The ball machine device (300) can supply a ball to the user based on information related to ball feeding among the above-mentioned posture measurement information. For example, if the information included in the posture measurement information entered by the user sets the goal as "right-handed, forehand, 4-second intervals, 10 practice repetitions, flat ball," the ball machine device can supply a tennis ball at a speed and angle corresponding to the information included in the posture measurement information.

[0140] Below, the camera operation is explained.

[0141] According to an embodiment of the present invention, the terminal (200) can transmit a control command to the camera (400) to record a video of the user's tennis play.

[0142] The camera can obtain a user posture image by recording a video of the user playing tennis based on the above control command.

[0143] At this time, the camera can be set to operate at the optimal angle and position to capture the user's entire posture.

[0144] The camera captures posture images continuously or at specific points during the user's practice session. The captured user posture images can be transmitted to a terminal (200) or server (100) for real-time or post-practice analysis.

[0145] Meanwhile, although this embodiment has described a system in which a camera is included as a separate device as an example, it is also possible for the user posture image to be captured by a camera provided in the terminal (200).

[0146] The camera can transmit the captured user posture image to the server (100) for similarity determination. Alternatively, the terminal (200) can transmit the posture measurement information and comparison target image information to the server for similarity determination.

[0147] Below, an example of implementation of a user guide screen for performing S310 to S330 in FIGS. 4 to 8 is described.

[0148] Figure 4 illustrates a detail measurement information input screen according to an embodiment of the present invention.

[0149] The terminal (200) can output a posture measurement information input screen for obtaining posture measurement information through a display.

[0150] Referring to FIG. 4, the posture measurement information input screen may include an option area (41) for inputting at least one piece of posture measurement information, an example screen (42) for showing an example of a coaching service provided by a tennis coaching program, and a navigator (43) for showing the progress of the coaching service, time information, and a screen transition icon.

[0151] The above option area (41) can support the user to select various posture measurement information related to the tennis posture the user wishes to practice.

[0152] As described above, the above posture measurement information may include at least one of the posture to be practiced, grip, ball launch interval, pitch quality, number of swings, ball drop location, and swing count voice information.

[0153] Users will be able to input various measurement information related to the tennis posture they wish to practice (e.g. posture information for forehand, backhand, serve, etc., grip information, ball launch interval, ball quality information, etc.) through the posture measurement information input screen.

[0154] According to an embodiment of the present invention, the example screen (42) may provide an example of analysis results comparing a comparison target image (e.g., a coach image) with a captured user posture image. This will allow the user to familiarize themselves in advance with the information they may receive when using the tennis coaching system.

[0155] Meanwhile, the above example screen (42) may also be able to output the user's previous coaching information.

[0156] The navigator (43) according to an embodiment of the present invention can output a time stamp indicating the time required to use the coaching service and a button for switching screens.

[0157] As above, if the user inputs the posture measurement information for which he or she wishes to receive coaching in the posture measurement information input screen, he or she can switch to the next screen using the screen change button of the navigator (43).

[0158] Figure 5 illustrates a comparison target image selection screen according to an embodiment of the present invention.

[0159] Referring to FIG. 5, the comparison target video selection screen may include at least one area among a comparison target video list area (51) in which a list of comparison target videos is listed, a comparison target video output area (52) for checking the comparison target video, a guide area (53) for obtaining a correct user posture image, and a navigator indicating the progress of the coaching service, time information, and screen transition icons.

[0160] The above navigator is omitted as it overlaps with the description in Fig. 4.

[0161] Specifically, the comparison target image list area (51) may include a list of at least one comparison target image. The user may select any one of the comparison target images included in the comparison target image list.

[0162] According to an embodiment of the present invention, the comparison target image output area (52) can output the selected comparison target image when any of the comparison target images included in the comparison target image list is selected. This allows the user to view the comparison target image he or she wishes to select.

[0163] According to an embodiment of the present invention, the guide area (53) can output precautions and brief coaching comments that the user should be aware of based on the posture measurement information obtained in FIG. 4.

[0164] Users can select a comparison video to use as a reference for analyzing and improving their practice posture through the comparison video selection screen.

[0165] The above comparison target videos may include demonstration videos by professional coaches or videos of the user's previous postures, which will allow the user to objectively compare and evaluate his or her current posture.

[0166] Figure 6 illustrates a progress guidance screen according to an embodiment of the present invention.

[0167] According to an embodiment of the present invention, the above-mentioned progress guidance screen is a screen for outputting a guidance screen for measuring user posture.

[0168] Referring to Fig. 6, an image may be output to guide that posture measurement is in progress along with a guidance message indicating the operation of the ball machine device in a specific area (63) of the progress guidance screen.

[0169] Figure 7 illustrates an example of a user posture image according to an embodiment of the present invention.

[0170] A camera according to an embodiment of the present invention can obtain a user posture image by receiving a control command from a terminal (200).

[0171] The above user posture video may include a user's posture captured within a tennis court.

[0172] The terminal or server can acquire multiple body joint points (71 to 75) of the user based on the user posture image. In addition, the terminal or server can determine a reference coordinate system for determining the reference of the multiple body joint points.

[0173] Through this, the server or terminal can obtain the coordinate values ​​of each of the plurality of body joint points based on the reference coordinate system.

[0174] The coordinate values ​​of each of the above joint points can be used to determine similarity with the image to be compared.

[0175] Meanwhile, the screens of FIGS. 4 to 7 functionally represent an embodiment for providing services of a tennis coaching system in which the plurality of areas are provided, and the location, information, and included screens of the areas are not limited to the above embodiment.

[0176] Additionally, it is not limited to embodiments such as selection methods for specific information, screen transition methods, icons, navigators, etc.

[0177] Figure 8 illustrates a waiting guidance screen output after acquiring a user posture image according to an embodiment of the present invention.

[0178] Referring to FIG. 8, the terminal may display a waiting guidance screen after a tennis ball is supplied based on posture measurement information. The waiting guidance screen may display text or an image (81) indicating that the acquisition of the user's posture image has been completed and indicating the time required to provide a tennis coaching service using artificial intelligence.

[0179] Additionally, the waiting guidance screen can display options (82, 83) to practice additional swings or re-take posture measurements during the user's waiting time.

[0180] Let's explain Figure 3 again.

[0181] According to an embodiment of the present invention, the terminal (200) can obtain coaching information based on the similarity between the user posture image and the comparison target image (S340).

[0182] And, the above coaching information can be output through a display (S350).

[0183] According to an embodiment of the present invention, the similarity determination between the user posture image and the comparison target image for generating coaching information may be performed on a server or a terminal. To this end, the server or terminal may each be equipped with a similarity determination model.

[0184] The following describes the operation of a server according to an embodiment of the present invention.

[0185] Figure 9 is a flowchart illustrating an operation method of a server for providing an AI coaching service according to an embodiment of the present invention.

[0186] According to an embodiment of the present invention, the server (100) can obtain a user's posture image and analyze the user's posture image to provide coaching information.

[0187] Referring to FIG. 9, in S901, the server can obtain a user posture image captured by a camera (400) of the user's posture.

[0188] The server can acquire user pose images from different angles of the same pose. The server can select images from among those taken from different angles of the same pose for analysis.

[0189] In S902, the server can obtain at least one body joint point from the user posture image.

[0190] Specifically, the server can obtain a skeleton of the user posture image using a computer vision algorithm.

[0191] Thereafter, the server can estimate the positions of multiple joints included in the acquired skeleton to obtain at least one body joint point. The body joint point may correspond to any one of the positions of the shoulder, elbow, hand, pelvis, knee, and ankle.

[0192] According to an embodiment of the present invention, the server can calculate a reference coordinate system for determining a reference of at least one body joint point (S903).

[0193] At this time, the above reference coordinate system can be based on the center of the body or a specific joint point, and can be a standard for quantitatively analyzing changes in the body's posture.

[0194] Specifically, the server can determine a reference coordinate system based on multiple body joint points acquired from a user posture image.

[0195] For example, the server can set the centers of the user's two shoulder joints as the reference points of the reference coordinate system. Alternatively, the server can determine the reference coordinate system based on the user's pelvic point.

[0196] According to an embodiment of the present invention, the server can obtain the coordinates of each of a plurality of body joint points based on a reference point of the reference coordinate system (S904).

[0197] Additionally, after setting the reference coordinate system, the server can calculate the relative positions and joint angles of multiple body joint points.

[0198] This allows for minimizing differences due to gender, physical condition, etc. when comparing the angles of the user's joint points with the comparison target image.

[0199] Additionally, by obtaining coordinates by setting a reference coordinate system, the problem of reduced precision can be prevented by reducing distortion due to the angle of view within the image.

[0200] In S905, the server can determine the similarity between the user posture image and the comparison target image based on the coordinates of the acquired joint points.

[0201] Below, the method for judging similarity is explained.

[0202] According to an embodiment of the present invention, the server can determine similarity based on the positions and angles of joint points included in each of the user posture image and the comparison target image.

[0203] At this time, the above similarity can be used to evaluate how much the user's posture matches the posture of the comparison target image.

[0204] According to an embodiment of the present invention, the server can obtain a first angle calculated as an angle between adjacent joints based on the coordinates of each joint point of the comparison target image.

[0205] Additionally, based on the coordinates of each joint point of the user posture image, a second angle can be obtained by calculating the angle between adjacent joints.

[0206] The server can calculate the similarity based on the difference between the first angle and the second angle.

[0207] In some embodiments, the server may acquire images previously captured and stored by the user. Specifically, the server may acquire and analyze images captured by the user or stored separately on the user's terminal.

[0208] According to an embodiment of the present invention, the server (100) can provide a service that provides guidance so that the user can take a certain posture during practice, or provides a coaching curriculum remotely or through AI.

[0209] Additionally, it is possible to distinguish actions in a video of a user actually playing a game, analyze the score for each action and consistency information according to game progress, and provide analysis results after the game.

[0210] According to an embodiment of the present invention, the server (100) can extract a certain posture motion and calculate the similarity even if the user practices or plays a game in which he or she performs random motions.

[0211] FIG. 10 is a conceptual diagram illustrating a method for a server according to an embodiment of the present invention to obtain coordinates of joint points of a human body.

[0212] According to an embodiment of the present invention, the server can obtain a plurality of body joint points from a user posture image to determine similarity in the user posture image.

[0213] At this time, the server (100) can train an artificial intelligence model to extract the user's joints or skeleton, or apply a model trained through transfer learning. The artificial intelligence model used by the server may utilize a deep learning-based algorithm, such as pose estimation. In addition, various artificial intelligences, such as openpose and posenet, may be utilized.

[0214] Referring to FIG. 10, the server (100) can extract the user's skeleton from the user posture image (1001) and obtain an image (1002) in which the skeleton is extracted with the wrist, elbow, shoulder, pelvis, knee, and ankle as main joint points.

[0215] According to an embodiment of the present invention, the server (100) can obtain the coordinates of each joint point from an image (1003) in which a reference coordinate system is determined.

[0216] Referring to FIG. 10, as shown in 1003, the server can determine a reference point (1004) based on each body joint point and obtain the coordinates of each joint. For example, the reference point (1004) may be the clavicle point, which is the center of the shoulder joint point. In addition, the server can increase the accuracy of the coordinates based on the reference point determined based on the body joint point as well as the clavicle point. At this time, the coordinates of each body joint point can be obtained in a 2D or 3D coordinate system.

[0217] The server (100) can obtain the angle formed by each body joint point with two adjacent body joint points based on the coordinates of each body joint point.

[0218] For example, the elbow angle can be obtained based on the hand, elbow, and shoulder joints. The server can obtain the angle of motion for each joint, such as the elbow, shoulder, pelvis, and knee.

[0219] The server (100) can calculate the similarity between the comparison target image and the user's posture image based on the acquired angle.

[0220] FIG. 11 is a flowchart illustrating a method for calculating the similarity of user posture images according to an embodiment of the present invention.

[0221] In the present invention, the user posture image includes multiple posture movements, and each posture movement image includes multiple sub-motions. The posture movements refer to tennis movements such as forehand, backhand, serve, volley, and slice. The multiple sub-motions may be configured differently for each tennis posture movement.

[0222] In an embodiment of the present invention, the server may divide the user posture image and the comparison target image into a plurality of images based on posture motion, divide each of the plurality of images divided based on the posture motion into detailed motions constituting the posture motion, and calculate the similarity based on the difference between the first angle acquired at each joint point of the detailed motion of the comparison target image and the second angle acquired at each joint point of the detailed motion of the user posture image.

[0223] For example, detailed movements for a forehand posture motion may include takeback, racket down, swing, follow through, etc. The server may determine the similarity of the user posture image by determining the similarity of each of the plurality of detailed movements.

[0224] According to an embodiment of the present invention, the server (100) can acquire a posture image of the user and segment the acquired posture image of the user by movement. (s1101)

[0225] Below, a method for segmenting a posture motion image from a user posture image is described.

[0226] More specifically, when the ball machine device is controlled by the terminal (200), the server can obtain a ball launch signal and segment the user posture image based on the ball launch signal.

[0227] Alternatively, the server (100) may segment the user's posture image based on the ball's position. In this case, the user image may be segmented based on the point at which the ball crosses the court. Alternatively, the image may be segmented by distinguishing movements based on the coordinates of specific body joint points.

[0228] The server can further divide the video divided into action units into detailed actions (s1102).

[0229] In the present invention, "sub-motion" may refer to the individual user movements that constitute a single motion. These sub-motions may be acquired based on one or more joint points. The server may set and store the minimum number of sub-motions required to analyze a single motion, and these sub-motions may be set differently for each type of motion.

[0230] The server can segment the user's posture video into frames. From the segmented frame image set, the server can obtain detailed motion-specific images necessary for comparison.

[0231] For example, if the user posture image is a forehand motion, the server can determine that it is a takeback if the coordinates of the wrist joint point have a first value in the x-axis direction or a second value in the y-axis direction relative to the reference point, or if they exist at a similar position in the y-axis direction.

[0232] Alternatively, the server may determine that the racket is down if the coordinates of the wrist joint point have a third value in the x-axis direction or a fourth value in the y-axis direction compared to the reference point.

[0233] Alternatively, the server may determine a swing if the coordinates of the wrist joint point have a fifth value in the x-axis direction or a sixth value in the y-axis direction.

[0234] Likewise, if the coordinates of the wrist joint point on the server have the 7th value in the x-axis direction or the 8th value in the y-axis direction, it can be determined as follow-through.

[0235] At this time, the absolute value of the first value, which is the x-axis coordinate value, may be a value smaller than the absolute value of the third value.

[0236] At this time, the absolute value of the 5th value, which is the x-axis coordinate value, may be greater than the absolute value of the 7th value.

[0237] Additionally, the second value, which is the y-axis coordinate value, may be a value greater than the fourth value.

[0238] Additionally, the 6th value, which is the y-axis coordinate value, may be a value smaller than the 8th value.

[0239] This is explained based on two-dimensional coordinates and is supported by Fig. 18.

[0240] The above coordinate values ​​are set based on an example forehand stroke and may be changed depending on the embodiment.

[0241] In this way, the server can identify each detailed movement based on the coordinates of the joint points.

[0242] The server can obtain the first angle, which is the angle value of the joint point required for comparison, by dividing the motion of the comparison target image into multiple detailed motions. (s1103)

[0243] The server can divide the comparison target image into multiple detailed motions and identify joint point information necessary for posture analysis from the acquired multiple detailed motions.

[0244] For example, in the case of a smash motion, the server can set the angle of the elbow joint point and the shoulder joint point of the dominant arm as the reference point for judgment, and in the case of a forehand motion, the server can set the elbow, shoulder, and hip angles as the reference points for judgment.

[0245] The server can obtain the angle values ​​of joint points for each detailed motion of the comparison target image as one data set (first angle).

[0246] The server stores the first angle, which is a data set for each video to be compared, and when the user selects the desired coaching data, the server can obtain the first angle of the selected coaching data and compare them.

[0247] The server can obtain the angles of body joint points from the acquired detailed motion images. The angles of body joint points are obtained for each motion, each detailed motion, and each joint point to form a data set, which is referred to as the second angle (s1104).

[0248] The server can obtain the difference by comparing the first and second angles obtained. (s1105)

[0249] According to an embodiment of the present invention, the server can compare the angles of identical body joint points at first and second angles. The server can compare the first and second angles in the same motion and the same detailed motion.

[0250] If the number of detailed movements of the comparison target video does not match the detailed movements of the user posture video, the server may match the detailed movements of the user posture video to the detailed movements of the comparison target video and may not judge the user posture videos that do not match.

[0251] The server can do the opposite if the user's posture image has more detailed movements.

[0252] In an embodiment of the present invention, detailed motions can be obtained as frame images. That is, an image with more frame images may not be judged as part of an image with fewer frame images.

[0253] The server can obtain the user's posture similarity based on the difference value obtained by comparing the matched first and second angles. (s1106)

[0254] According to an embodiment of the present invention, the server can calculate the similarity of each joint point based on a criterion set for the joint point and synthesize the similarity to calculate the overall similarity.

[0255] More specifically, the server can set the angle range for similarity assessment for each joint point. This range can vary for each joint point, and even for the same joint point, the angle range can vary depending on the movement. For example, the angle at the elbow joint in a forehand shot may have a narrower similarity assessment range than the angle at the elbow joint in a serve, and thus may be evaluated more strictly.

[0256] Additionally, the server can set different angle ranges based on gender, height, etc., and can set different angle ranges for different joint points within a single movement.

[0257] According to an embodiment of the present invention, the server can calculate this similarity as a score for each part and can calculate it numerically as an overall motion score.

[0258] Figure 11 illustrates a typical case where the comparison target image is a specific image. However, even if the user does not select a specific image, the server can store the reference joint point standard for each movement and evaluate and determine the user's posture based on this. In this case, in s1101, the comparison target image is not acquired, and only the angle of each detailed movement that serves as the reference is used to obtain the first angle, and the user's posture can be evaluated by comparing this with the second angle. In this case, the similarity can be determined even if the user does not select a specific image.

[0259] Meanwhile, if the first angle of the comparison target image is an already stored value, s1103 can be performed before s1101 and s1102, and if the user wants to analyze only one movement, s1101 can be omitted, and so on. Various embodiments can be configured and used.

[0260] In an embodiment of the present invention, the server can generate coaching information based on the similarity. Based on the similarity for each posture, the server can calculate the posture with the highest and lowest similarity as the worst and best posture motions, respectively, and generate coaching information based on these. In this case, the server can generate the coaching information by including a comparison target image.

[0261] The server can determine whether a user has an incorrect posture based on the above similarity. The server can identify posture movements or detailed movements whose similarity is determined to be below a set value, and generate coaching information for the incorrect posture based on the posture movements or detailed movements.

[0262] If the server identifies a user's posture movements or detailed movements that are judged to be incorrect more than a preset number, the server may determine that the user frequently makes incorrect posture movements or detailed movements and generate coaching information. The server may also acquire previous posture movements determined to be below a preset value from among the stored user posture images and generate coaching information for them. Alternatively, the server may identify frequently incorrect posture movements and generate coaching information for them.

[0263] In an embodiment of the present invention, the server may provide the user with coaching information by generating a user's posture score and growth graph based on similarity, or by providing timelapse video and detailed movements in a side-scroll format.

[0264] More specifically, the server can determine the score for each posture movement by assigning scores to detailed movements with a similarity of 80% or more. The server can store the scores for each posture movement and generate a graph showing the trend of scores for each posture movement, which can then be used as coaching information.

[0265] The server can generate coaching information visualizing the similarity obtained in Figure 11 for the user. This is further explained in Figure 12 below.

[0266] FIG. 12 is a diagram illustrating an example of posture similarity provided by a server to a user according to an embodiment of the present invention.

[0267] The server can compare a user posture image (1201) with a comparison target image (1202) in the same detailed motion to determine the similarity of the user's posture. If the server determines the user's posture based on the angles of each joint point, it can generate coaching information (1203) to provide the user with this information.

[0268] The user posture image (1201) represents a detailed movement of the user posture image, in which the joint positions are estimated and displayed in a skeleton form. More specifically, the server can display the acquired body joint points of the wrist, elbow, shoulder, pelvis, knee, and ankle in the user posture image.

[0269] The comparison target image (1202) is the coaching data used as the comparison target image in this example. In this drawing, the user posture image (1201) and the comparison target image (1202) display the same body joint points, and both movements correspond to a sub or smash movement.

[0270] The server (100) can obtain angles of the elbow, shoulder, pelvis, and knee joints among the wrist, elbow, shoulder, pelvis, knee, and ankle joints. The server can compare the angles of each joint point between the user posture image (1201) and the comparison target image (1202).

[0271] The server can visualize the difference in angles and reflect them in the user's posture image to generate coaching information (1203) that analyzes the user's posture. The server can compare each joint point of the user with the angles of the joint points in the comparison image, visualize the difference in angles and their directions, and provide feedback to the user. The visualization of the difference in angles and directions can be expressed in a form such as 1204.

[0272] This way, we can communicate more concisely and clearly not only the overall score but also the areas where feedback is needed for each body part.

[0273] FIG. 13 is a flowchart illustrating a method for determining consistency information from a user posture image according to an embodiment of the present invention.

[0274] According to an embodiment of the present invention, the server can divide the user's posture image into motions and provide the user with consistency information indicating the degree to which the user's posture is maintained as the user performs the motions.

[0275] In the present invention, consistency information refers to information indicating the degree to which a user's posture is maintained when repeatedly performing a posture. According to an embodiment of the present invention, the server can determine consistency information to enable a user who repeatedly performs the same movement to maintain a consistent posture.

[0276] According to an embodiment of the present invention, the server obtains a plurality of user posture motion images segmented based on posture motion from the user posture image, identifies a first user posture motion image as a reference among the plurality of user posture motion images, and calculates the degree of similarity between the first user posture motion image and other user posture motion images to determine consistency information.

[0277] In the present invention, the server can provide consistency information based on whether the angle of the user's joint point is maintained, or can provide consistency information based on whether the coordinates of the user's joint point are maintained.

[0278] The motion-specific segmentation of the user posture image of the present invention has been described above, so it is omitted. (s1301~1302)

[0279] The server (100) can select a reference image from among the segmented motion-based images. The server can use the first acquired user posture image as the reference. Meanwhile, if the first user posture image is lower than the comparison target image or a set reference, the server can use the second user posture motion image as the reference.

[0280] Below, a method for calculating consistency information is described using the first user posture motion video as an example.

[0281] The server (100) can compare the coordinates or angles of the second user posture motion image and the third user posture motion image with the first user posture image. At this time, the coordinates or angles may be set differently for each motion, and may generally be based on the body joint point with the most movement.

[0282] More specifically, when a user performs a forehand motion with his right hand, the server can calculate the x-coordinate of the right wrist and compare it with the first user image and the remaining images.

[0283] The server (100) can calculate a first similarity by comparing a first user posture motion image with a second user posture motion image. The server can calculate a second similarity by comparing a first user posture motion image with a third user posture motion image. If the motion is divided into three or more, each user posture image can be compared with the first user posture image to calculate a similarity for each image divided for each motion. Here, the method for calculating the similarity is omitted because it has been described in FIGS. 11 and 12.

[0284] Additionally, the server can derive consistency information based on the similarity obtained from each segmented image according to the user's actions. Based on each similarity, the server can generate a graph or visualization of the change in similarity and provide it to the user.

[0285] The description of Fig. 13 described above is a schematic representation of consistency information when the user posture image consists of three movements. Depending on the user posture image, the description of this drawing may need to be applied differently. In addition, unlike this description, a consistency graph may be directly calculated using coordinates for each detailed movement without calculating similarity. The consistency graph is described later in Figs. 18 and 19.

[0286] FIG. 14 is a flowchart illustrating a method by which a server determines a user posture image for determining similarity according to an embodiment of the present invention.

[0287] According to an embodiment of the present invention, the server (100) can acquire a user posture image from a camera (200) to determine similarity. The server can acquire and analyze a single user posture image, or can acquire multiple user posture images via multiple cameras. More specifically, the server can determine a camera to capture images from among the multiple cameras, or acquire multiple user posture images and then determine an image to analyze from among them.

[0288] According to an embodiment of the present invention, when the ball machine device (300) is controlled and the ball is fed in response to the movement entered in the posture measurement information in the terminal (200), the server (100) can obtain the posture measurement information entered by the user from the terminal. (s1411)

[0289] The user's posture measurement information may include at least one of the user's practice posture motion information, such as forehand, backhand, serve, and slice. The camera (400) can capture the user from two or more positions, and the terminal (200) can select a camera to capture the image based on the user's posture measurement information. (s1412)

[0290] According to an embodiment of the present invention, if the user inputs a forehand as the posture measurement information, the terminal (200) may determine a camera located on the user's right side as the recording camera. Alternatively, if the user inputs a backhand as the posture measurement information, the terminal (200) may determine a camera located on the user's left side as the recording camera.

[0291] The above decision method can select a shooting camera to be analyzed based on a set value when the user's actions are obtained in advance from the terminal (200).

[0292] Even if the ball machine device (300) is not controlled according to the user's actions on the terminal (200), the server (100) can determine an analysis image to obtain an appropriate user posture.

[0293] According to an embodiment of the present invention, two or more cameras may be installed in an outdoor practice range or an outdoor stadium. The server (100) may determine a primary analysis camera among the images acquired from the two or more cameras to determine an analysis image.

[0294] According to an embodiment of the present invention, when a practice / game starts, the server (100) can acquire images from two or more cameras (s1421), identify factors for determining images to be analyzed (s1422), and determine images to be analyzed among the acquired images (s1423).

[0295] According to an embodiment of the present invention, the server (100) can identify the user's main hand direction, the user's swing, the user's position within the court, the user's motion, the user's torso direction, etc. from the user's posture image to determine an appropriate camera for identifying the user's skeleton. In this way, the terminal (200) can select and analyze images even in random motions or actual games where the motion is determined and the posture is not analyzed.

[0296] According to an embodiment of the present invention, the server (100) or terminal (200) can adjust the direction and angle of the camera. While described below as the operation of the server (100), this can also be performed by the terminal (200). The server (100) can identify the user's location and control whether the camera should be adjusted. If the user is not located within the area captured by the camera or the user's posture is not sufficiently identified in the image, the direction and angle of the camera can be adjusted.

[0297] According to an embodiment of the present invention, a user's posture image can be analyzed using all images acquired from two or more cameras. More specifically, the server can utilize the differences between the images acquired from each camera to obtain the 3D coordinates of an object. In this case, the coordinate system of the user's body joints can be obtained in 3D to calculate similarity.

[0298] Meanwhile, according to various embodiments, the terminal (200) may select any image among the types of comparison target images as a comparison target image if a separate comparison target image is not determined by the user.

[0299] Specifically, the server may determine an image selected by user input from among the coach image and the user's previous user posture images as the comparison target image. Alternatively, the server may determine one of the following as the comparison target image: the first user posture image with the highest similarity to the comparison target image among the user's previous user posture images, the first image among the user's posture images, and the previously used comparison target image.

[0300] Below, the coaching information provided by the tennis coaching system is described in FIGS. 15 to 19.

[0301] According to an embodiment of the present invention, the coaching information can be generated based on a user posture image and a comparison target image.

[0302] The server can contribute to improving the user's tennis skills by providing feedback on the user's tennis posture, similarity with the comparison video, and posture consistency information through coaching information.

[0303] According to an embodiment of the present invention, the coaching information may include at least one of a similarity calculation result between the user posture image and the comparison target image, posture measurement information, a record of the user posture image, an option of a comparison screen, and feedback for the user posture image.

[0304] The terminal can output coaching information obtained from the server.

[0305] Figure 15 illustrates a coaching screen for outputting coaching information according to an embodiment of the present invention.

[0306] Referring to FIG. 15, the terminal (200) can output a coaching screen including at least one area among a comparison area (1510) that displays a feedback image for the user posture image, an information display area (1520) that displays preset posture measurement information, a recording area (1530) that displays a record of the user posture image, an option area (1540) for setting options of a comparison screen output to the comparison area (1510), and a feedback area (1550) that provides feedback for the user posture image.

[0307] The terminal (200) can output a user posture image through the comparison area (1510). At this time, the terminal (200) can output an option set in the option area (1540) together with the user posture image.

[0308] For example, if the options set in the option area (1540) are 'swing trajectory display' and 'next section automatic playback on', the terminal (200) can automatically play the next section of the user posture image while displaying the swing trajectory of the user posture image in the comparison area (1510).

[0309] Based on various embodiments, options included in the options area (1540) may include options for displaying comparison target images (see FIG. 12) and comparison and joint data (see FIG. 7).

[0310] In addition, the terminal (200) can output posture measurement information and video speed options through the information display area (1520). In addition, the recording area (1530) can distinguish multiple tennis posture movements (e.g., forehand, backhand, sub-slide, volley, etc.) included in the entire user posture video.

[0311] Additionally, the terminal (200) can provide detailed coaching of the posture movements practiced by the user through the feedback area (1550).

[0312] FIG. 16 and FIG. 17 illustrate an embodiment of a feedback area (1550) for outputting coaching information according to an embodiment of the present invention.

[0313] Referring to FIGS. 16 and 17, the feedback area may include at least one of comprehensive information (1551) including posture measurement information, detailed coaching information (1552) including feedback for each detailed motion (e.g., take back, racket down, swing, etc.) of the posture motion practiced by the user, posture accuracy information (1553) indicating the result of comparing the comparison target image and the user posture image, posture consistency information (1554) indicating consistency for the posture motion practiced by the user (e.g., any one of forehand, backhand, volley, and serve), and consistency graphs (1556, 1557).

[0314] Specifically, the detailed coaching information (1552) may be determined based on the similarity of each of the multiple detailed movements included in each of the user posture image and the comparison target image.

[0315] More specifically, detailed coaching information (1552) may output the number of repetitions and successes for a posture practiced by the user. The success rate may indicate a case where the similarity between the posture practiced by the user and the comparison target image exceeds a predetermined value.

[0316] Detailed coaching information may include feedback on recurring posture errors. This feedback may include advice on improving posture, points to note to improve posture accuracy, and areas of inconsistency identified through practice.

[0317] For example, feedback for each of the above detailed movements may include 'takeback is late', 'lower left hand', and 'right arm needs to be extended more when taking back'.

[0318] Additionally, the plurality of detailed movements can be determined based on the coordinate values ​​of wrist joint points obtained from the user posture image.

[0319] Referring to Fig. 17, the consistency graph shows the coordinate values ​​of wrist joint points obtained from the user posture image and the time information corresponding to the coordinate values, which can be used as an important indicator for determining the user's posture consistency.

[0320] For example, the coordinate value of the wrist viewpoint may be the height coordinate value of the wrist joint point.

[0321] According to an embodiment of the present invention, the consistency graph may include a first consistency graph (1556) representing the user's posture consistency and a second consistency graph (1557) representing the posture consistency of the comparison target image.

[0322] According to an embodiment of the present invention, the first consistency graph (1556) may represent time information corresponding to the coordinate values ​​of the wrist joint points obtained from the user posture image.

[0323] Additionally, the second consistency graph (1557) may represent time information corresponding to the coordinate values ​​of the wrist joint points obtained from the comparison target image, such as the coach image.

[0324] Figures 18 and 19 illustrate examples of consistency graphs based on a forehand motion according to an embodiment of the present invention.

[0325] Referring to FIGS. 18 and 19, for example, when a server analyzes a forehand stroke in tennis, the forehand stroke can be broken down into four detailed steps: takeback (1811), racket down (1812), swing (1813), and follow-through (1814).

[0326] As in the first consistency graph (1556) of FIG. 19, the take back (1811) may correspond to the first point (1901), the racket down (1812) may correspond to the second point (1902), the swing (1813) may correspond to the third point (1903), and the follow through (1814) may correspond to the fourth point (1904).

[0327] The server can generate a consistency graph by displaying the wrist joint movement at each step as a separate mapping (plot) on the consistency graph. This allows for a clear understanding of each detailed movement performed by the user.

[0328] Similarly, the server can generate time information corresponding to the coordinate values ​​of the wrist joint points obtained from the comparison target image as a second consistency graph (1557).

[0329] The take back of the comparison video may correspond to the first point (1911), the racket down may correspond to the second point (1912), the swing may correspond to the third point (1913), and the follow through may correspond to the fourth point (1914).

[0330] According to an embodiment of the present invention, the server can precisely analyze a user's wrist movement patterns over time through the consistency graph. This can be used as a key factor in evaluating the accuracy and consistency of a tennis stroke.

[0331] Meanwhile, according to an embodiment, the tennis coaching system (1) may further include a user terminal (500) that receives the coaching information from the server and outputs the received coaching information.

[0332] As described in FIG. 15, the user terminal (500) may output at least one of a comparison area showing a result of calculating a similarity between the user posture image and the coach image, an information display area including posture measurement information, a recording area recording the user posture image, an option area setting an option of the comparison screen, and a feedback area providing feedback on the user posture image.

[0333] According to an embodiment of the present invention, the terminal may provide an interface to enable a user to save an image of his or her posture measurement results or share it on a social network service (SNS).

[0334] This will allow users to personally record the results of their posture analysis obtained through a tennis coaching system that provides AI coaching functions, or share their progress through interactions with friends.

[0335] For example, the terminal (200) may provide an interface that allows the user to save the resulting image or upload it directly to a selected social networking service (SNS) platform. An application within the terminal may allow the user to easily select the image and set sharing options.

[0336] Figure 20 illustrates the configuration of a terminal, server, and user terminal of a tennis coaching system (1) according to an embodiment of the present invention.

[0337] Hereinafter, for convenience of explanation, the above devices are referred to as computing devices that provide coaching services.

[0338] A computing device according to an embodiment of the present invention may include a communication unit (110), an input unit (120), a sensing unit (130), a display (140), a memory (150), and a processor (160).

[0339] The communication unit (110) can transmit and receive data with external devices such as other electronic devices or servers using wired or wireless communication technology. For example, the communication unit (110) can transmit and receive sensor information, user input, learning models, control signals, etc. with external devices.

[0340] At this time, the communication technologies used by the communication unit (110) include GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), LTE (Long Term Evolution), 5G, WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Bluetooth (Bluetooth), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), ZigBee, NFC (Near Field Communication), etc.

[0341] The input unit (120) can obtain various types of data.

[0342] At this time, the input unit (120) may include a camera for inputting a video signal, a microphone for receiving an audio signal, a user input unit for receiving information from a user, etc. Here, the camera or microphone may be treated as a sensor, and a signal obtained from the camera or microphone may be referred to as sensing data or sensor information.

[0343] The sensing unit (130) can obtain at least one of terminal internal information, terminal surrounding environment information, and user information by using various sensors.

[0344] The display (140) can generate output related to visual, auditory or tactile sensations.

[0345] At this time, the display (140) may include a display unit that outputs visual information, a speaker that outputs auditory information, a haptic module that outputs tactile information, etc.

[0346] The memory (150) can store data that supports various functions of the computing device. For example, the memory (150) can store input data obtained from the input unit (120) or various data obtained from a connected device.

[0347] A processor configured to perform a specific operation may be configured to perform the specific operation, and may be designed in hardware as well as structured through programming to perform the specific operation.

[0348] The processor (160) can determine at least one executable operation of the computing device. Then, the processor (160) can control components of the computing device to perform the determined operation.

[0349] To this end, the processor (160) can request, retrieve, receive or utilize data from the memory (150) and control components of the computing device to execute an operation that is predicted or determined to be desirable among at least one executable operation.

[0350] At this time, if the processor (160) requires connection of an external device to perform a determined operation, it can generate a control signal for controlling the external device and transmit the generated control signal to the external device.

[0351] The processor (160) may control at least some of the components of the computing device to run an application program stored in the memory (150). Furthermore, the processor (160) may also operate two or more of the components included in the computing device in combination to run the application program.

[0352] Those skilled in the art will appreciate that the various exemplary logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented by electronic hardware, various forms of program or design code (referred to herein for convenience as software), or a combination of both.

Claims

1. On the server providing tennis coaching services, a communication unit for obtaining posture measurement information; and A processor for acquiring a user posture image and analyzing the acquired user posture image; The above processor Calculate the similarity between the user posture image and the comparison target image, Providing coaching information based on the above similarity, Server.

2. In paragraph 1, The above comparison target image includes at least one of a coach image and a previous user posture image of the user. Server 3. In paragraph 1, The above processor, Obtain multiple body joint points from the above user posture image, Determine a reference coordinate system for determining the reference of the above multiple body joint points, Obtain coordinate values ​​of the plurality of body joint points based on the above reference coordinate system, A server that obtains similarity based on the above coordinate values.

4. In paragraph 1, The above processor, Based on the coordinate values ​​of each joint point of the image to be compared, a first angle is obtained by calculating the angle between adjacent joints, Based on the coordinate values ​​of each joint point of the user posture image, a second angle is obtained by calculating the angle between adjacent joints, A server that calculates the similarity based on the difference between the first angle and the second angle.

5. In paragraph 1, The above processor, Splitting the user posture image and the comparison target image into multiple images based on posture movements, Each of the plurality of images segmented based on the above-mentioned posture motion is segmented into detailed motions constituting the above-mentioned posture motion, A server that calculates the similarity based on the difference between the first angle obtained at each joint point of the detailed motion of the comparison target image and the second angle obtained at each joint point of the detailed motion of the user posture image.

6. In paragraph 1, The above processor, Obtaining multiple user posture motion images segmented based on posture motion from the above user posture image, Identifying a first user posture motion image that serves as a reference among the plurality of user posture motion images, Calculate the similarity between the first user posture motion image and other user posture motion images, A server that calculates consistency information based on each of the above similarities.

7. In paragraph 5 The above processor, The similarity is determined for each of the above detailed actions, If the above similarity is determined to be below the set value, the image of the detailed action is identified, A server that generates coaching information based on the identified image.

8. In a tennis coaching system using artificial intelligence, A terminal that obtains user posture measurement information and transmits the posture measurement information to a server; and Obtain user posture images, Calculate the similarity between the user posture image and the comparison target image Including a server that provides coaching information based on the above similarity, Tennis coaching system.

9. In paragraph 8, The above comparison target video is, Containing at least one of a coach video and a previous user posture video of the user; Tennis coaching system.

10. In paragraph 8, The above comparison target video is, Among the previous user posture images of the user, the first user posture image with the highest similarity to the comparison target image, the first image among the user's posture images, and Contains any one of the previously used comparison images, Tennis coaching system.

11. In paragraph 8, The above server, Obtain multiple body joint points from the above user posture image, Determine a reference coordinate system for determining the reference of the above multiple body joint points, Obtain the coordinate values ​​of each of the plurality of body joint points based on the above reference coordinate system, Obtaining similarity based on the above coordinate values, Tennis coaching system.

12. In paragraph 8, The above server, Based on the coordinate values ​​of each joint point of the image to be compared, the first angle and the angle between adjacent joints are calculated. Based on the coordinate values ​​of each joint point of the user posture image, a second angle is obtained by calculating the angle between adjacent joints, Calculating the user's motion similarity based on the difference between the first angle and the second angle. Tennis coaching system.

13. In paragraph 8, The above terminal, Posture measurement information input screen for obtaining the above-mentioned posture measurement information, A comparison target image selection screen for determining the comparison target image and Output at least one screen among the guidance screens that provide guidance while acquiring the above-mentioned detailed measurement information, Transmitting the posture measurement information obtained based on user input and information on the comparison target image to the server. Tennis coaching system.

14. In paragraph 8, The above detailed measurement information is Containing at least one of posture information to practice, grip information, ball launch interval information, ball quality information, number of swings information, ball drop location information, and swing guide information. Tennis Coaching System 15. In paragraph 8, The above coaching information is Including at least one of the results of calculating the similarity between the user posture image and the comparison target image, posture measurement information, the user posture image record, options of the comparison screen, and feedback information for the user posture image. Tennis Coaching System 16. In paragraph 8, Further comprising a user terminal that receives the coaching information from the server and outputs the received coaching information, The above user terminal, Outputting at least one of a comparison area showing the similarity comparison result between the user posture image and the comparison target image, an information display area including posture measurement information, a recording area recording the user posture image, an option area setting options of the comparison screen, and a feedback area providing feedback on the user posture image. Tennis Coaching System 17. In paragraph 8, The above coaching information includes at least one of comprehensive information, detailed coaching information, posture accuracy information, posture consistency information, and consistency graph information. The above detailed coaching information is determined based on the similarity of each of the multiple detailed movements included in each of the user posture images and the comparison target images. Tennis coaching system.

18. In paragraph 17, The above multiple detailed movements are determined based on the coordinate values ​​of the wrist joint points obtained from the user posture image. Tennis Coaching System 19. In paragraph 17, A tennis coaching system in which the above consistency graph represents time information corresponding to the coordinate values ​​of the wrist joint point obtained from the user posture image.

Images